Children's exposure to aggression in the home is widespread in the United States, with several million either experiencing or witnessing violence. While serious physical injuries and death can result from family violence, the psychological injuries sustained are likely to be profound, long- lasting and more widespread. Among these psychological sequelae is the risk that children from violent homes will develop into perpetrators themselves. Although the idea that aggression is transmitted across generations is compelling, empirical support for intergenerational transmission remains scarce and inconclusive. The few studies that have been conducted rely on retrospective childhood reports that are subject to error due to memory failure or reconstruction. This lack of prospective, longitudinal work leaves a serious gap in our understanding of how aggression is maintained or disrupted across generations, and what factors might serve to buffer children. The goal of our proposed research is to follow for three years a sample of 375 women and children (Hispanic, Anglo-American and Native American), about half of whom were originally screened for the presence of wife-battering. Our proposed study is the first prospective study of children of battered women, and the first to include dating violence among other forms of aggression in adolescents at risk' Our study is also among the first to include a large sample of Hispanics, and should contribute to our understanding of how culture contributes to the expression of family conflict and psychopathology. The mean age of the children at our initial assessment was 9.3 years; the mean age at the outset of our proposed research will be approximately 14. The availability of this sample offers a unique window of opportunity by which to examine the transition to adolescence of children from violent and comparison homes. We propose to interview these mothers and one of their children three times over a five year period, assessing changes in family violence or structure, mother and child psychopathology, and social relationships within and outside of the family for the presence of both conflict and support We have organized our measures into two conceptual- categories tapping children's internal working models (social cognition, relationships, close family relationships) and social ecology (peer networks, sociodemographics, neighborhood cohesion). These socio- emotional, social cognitive and ecological factors might serve to mediate or buffer aggression in adolescence. We hope that by addressing the multifaceted nature of aggression, both within and outside the family, and by examining how this family aggression is translated into children's mental models of relationships, we will better understand the mechanisms involved in the intergenerational transmission of aggression. Our findings should inform intervention efforts for agencies providing services both to battered women and to maltreated children, as well as prevention efforts for violence against women launched in community settings and schools.